The Film Finder was first a university project which was proposed to help look into different ways of data mining data from user's interaction with the website. The main goal of the project was to create a website front end which allows the users to interact with the data in the database. This mean it had to be an easy and user friendly web interface. This was achieved by creating a few designs and the current site you're looking at was the final winning design. This meant that the remainder of the time was spend on creating the backend database and algorithm to generate the recommendation. During the university project the algorithm went through 4 updated version. This was done with help of Perforce Software tools. These tools made it easy to create the website and keep a history of all the changes that have happen over the duration of the project.
The idea of the website was create by myself based on the idea that I enjoyed watching films and couldn't think of what I should watch next. This is where the drive for getting this project created and finished was coming from. Once the project was about half way through the process I noticed that there was a real chance this could be a good website to publish to the outside world. After the university project was over, I started to create a new database sourced from different data source. I also re-design some of the website to make it comply with the W3C standards.
The idea for the user is to input a movie title they enjoyed watching in the past and the website will return all the information about that movie. The user can checked this is the movie they were intending to look for. Once they have confirmed this was the movie they were looking for, they can look at the recommendation for this movie. They can get more information for each of the recommendation by clicking on the poster of the movie. The decision to use poster over text was because most people are visual guided rather than reading lots of text. By just click on a movie the user has helped to increase the recommendation higher up the list. If the user thinks it's a bad film they can click the bad recommendation button which will decrease it down the list or remove it completely.
The basic idea of the recommendation is using the basic information for the movie and putting this through the algorithm that is create in-house, and generated a basic layer of recommendation. There is no recommendation system out there that can create the best recommendation for every user. This is then built upon the user interaction with the website by them voting for the movie recommendation.